Rain, Risk and Resilience

Initial reflections on storm Daniel’s flood disasters

At the beginning of September 2023, a low-pressure system formed in the eastern part of the Mediterranean basin, resulting in a storm (named “Daniel”) that would become the deadliest and costliest Mediterranean tropical-like cyclone (informally referred to as medicanes) ever recorded. The resulting floods and their unprecedented impacts in Libya, Greece, Bulgaria and Turkey put water systems under immense stress, brought systematic shortcomings as well as profound complexities involved in dealing with natural hazards to the forefront and put existing risk management and adaptation approaches in question.

During storm Daniel, in the village of Zagora in Greece, 754 mm of rain was recorded in less than 24 hours. To put this into perspective, in the Netherlands, an average of 821 mm of rain fell throughout the whole of 2022. In a nearby village, Makrinitsa, a cumulative 1235 mm of rain has been recorded so far in September, making it the highest monthly rainfall ever recorded in Europe. The immediate aftermaths of this unprecedented rainfall were tens of thousands of casualties (along with the perishing of hundreds of thousands of animals), tens of thousands of people displaced, the overwhelming of two dams in Libya and the collapse of multiple bridges, severe damage to critical infrastructure and economic losses at the scale of billions of Euros in Greece alone. However, as time goes on, further cascading impacts are anticipated, including health risks (associated with contaminated water, mental health, disrupted healthcare systems etc.), far-reaching impacts to the water-energy-food nexus (e.g. as a result of immense stress in agriculture), economic and political instability and numerous other impacts.

As relief and recovery efforts take place, we need to reconsider the way we approach and prepare for natural hazards in general, in order to enhance the resilience of communities in the face of such events and avoid future disasters.

Precipitation accumulation values from NASA’s IMERG multi-satellite product from Sept. 3 - 7, 2023, showing the rainfall that ensued from storm Daniel. Source: NASA IMERG

Precipitation accumulation values from NASA’s IMERG multi-satellite product from Sept. 3 – 7, 2023, showing the rainfall that ensued from storm Daniel. Source: NASA IMERG

Changing environmental patterns and deep uncertainty

Storm Daniel was undoubtedly an extremely intense and rare event, with preliminary estimates suggesting a return period of (well) above 1000-years. However, the return period cannot even be estimated for some places on the plain, as the rain gauging stations stopped functioning altogether, after they were flooded. In simple terms, a return period tells us how likely it is for a hazard event, to reach (or surpass) a certain intensity in any given year. For example, a 1-in-100 year event means that there is 1% chance of experiencing a similar or more intense event within one year, rather than that it happens only once in 100 years.

However, just three years earlier (on September 2020) another high-impact medicane, storm Ianos, impacted the eastern Mediterranean leading to the loss of 5 people, 13000 livestock animals and an estimated € 700 million in damages and € 31 million in insured losses. The return-period of that event was estimated at around 1000 years, essentially meaning that there should have been a nearly zero percent chance of such storms happening at such a short interval, putting our existing statistical estimates into question. And climate change may play a role in this.

In a recent preliminary report on Storm Daniel’s characteristics and impacts, it is stated that “preliminary results based on weather analogues, show that human induced climate change has increased the intensity of such storms in the recent years”. This observation is also in line with IPCC’s assessment of climate change impacts on the Mediterranean region, that warns of intensification of rainfall extremes in the northern part of the region. Regarding medicanes in particular, our current knowledge forecasts a frequency decrease in parallel to an intensity increase, but our models entail too many uncertainties to draw sound conclusions. However, on a broader level flooding may become much more frequent and intense in the future according to recent studies.

At first, these observations may indicate a need to revise existing risk assessments and strategies, such as the flood risk assessments and management plans that were developed under EU’s Floods Directive (2007/60/EC), by considering modified frequencies and intensities of extreme weather events due to climate change. However, at a deeper level we may need to rethink the way we approach traditional engineering design and risk management all together. As the uncertainties entailed in current climate models and those that arise from unknown climate tipping points and mechanisms make it hard to adequately predict future conditions, we may need a shift from a probability-based mindset to a plausibility-based one, as recently explained in

Van Thienen, P., Chatzistefanou, G.A., Makropoulos, C. & Vamvakeridou-Lyroudia, L. Maintaining our water supply systems in a conceivable societal collapse. The 19th Computing and Control for the Water Industry Conference, 4-7 September 2023, De Montfort University, Leicester, UK.

and past reports.

Water systems and critical infrastructure in focus

The immense flooding that ensued from storm Daniel, imposed extreme strains on the infrastructure of the affected areas, in some cases well beyond the design considerations with which they were built. Apart from the overwhelming of the stormwater and drainage networks, which had devastating consequences in Libya as two dams were breached, power outages, transportation disruptions (including due to the collapse of several bridges) and water supply interruptions were experienced. Preliminary data collected by the author on the wider area of Volos in Greece, reveal extensive damage to the water distribution network (due to the destruction of several water mains amongst others), leading to water shortages, the operation of the network in an intermittent supply state, as well as significant water quality problems. At the time of writing this article, almost two weeks after the receding of the extreme rainfall, the normal operation of the water network still hasn’t been restored.

People queue to receive bottled drinking water in Volos, amidst severe drinking water supply disruption in the aftermath of storm Daniel. Source: InTime News

People queue to receive bottled drinking water in Volos, amidst severe drinking water supply disruption in the aftermath of storm Daniel. Source: InTime News

Some weeks earlier, at the end of July, the same water distribution network (around Volos) was disrupted by wildfires, bringing the issue of multi-hazards and consecutive disasters in the foreground. As climate risks intensify, infrastructure may become exposed to different hazards at shorter intervals than they can recover, leading to magnified impacts. This issue entails several challenges and intricacies, that we don’t fully understand yet. Ongoing research at KWR, such as the EU funded ARSINOE project (2021-2025), EU funded IMPETUS project (2021-2025) and the EU funded NATALIE project, (2023-2028), aim to advance our understanding and scientific insights on enhancing resilience to climate change, potentially resulting in pathways and aid in coping with it.

Drone footage of the flood extents of the Pineios river near Larissa, Greece after the receding of the storm.

Drone footage of the flood extents of the Pineios river near Larissa, Greece after the receding of the storm.

Apart from the apparent engineering aspect, the widespread infrastructure impacts revealed several other dimensions of infrastructure resilience. For example, the dam breaks in Libya appear to have been caused due to their deteriorating state, although details are still emerging, which has deep political and economical dimensions. On a smaller scale, considerable challenges were posed to the local water utility of Volos, regarding the governance and communication aspects, as there is reported confusion in the local population regarding the safety and availability of drinking water. Additionally, the complex issue of infrastructure interdependencies became apparent. Modern infrastructure systems (such as electricity, water, transportation or telecommunication networks) are closely interlinked, interacting with and affecting each other. As a result, failures in one system may propagate to others. For example, in Thessaly, Greece essential machinery was reported damaged in the aftermath of prolonged power outages.

A collapsed bridge near the village of Kala Nera, Greece on September 7, 2023. Source: Reuters

A collapsed bridge near the village of Kala Nera, Greece on September 7, 2023. Source: Reuters

It becomes therefore apparent, that building infrastructure resilience requires a multidimensional approach, considering infrastructure interactions, governance aspects and the wider environment (natural, political, economical, etc.) in which they operate in, amongst other. Research regarding these issues may complement a holistic strategy to tackle this admittedly complex challenge.

Floods happen – disasters may be prevented

The disasters that followed storm Daniel revealed the complex processes that natural hazard mitigation entails, uncovered systematic shortcomings in response mechanisms, and put the way we live with water in general, into perspective. Flood risk remains a global challenge, that factors like rapid urbanization, poverty and a changing climate amplify. In order to cope with this risk, we must drastically rethink the way we interact with water (and the extent to which we have “imposed” ourselves to the natural ways of water) and try to create a mature society that recognizes floods as a continuing, but manageable threat. To that extent, the deployment of nature-based solutions and the restoration of natural waterways and floodplains, along with the education of communities regarding flood risk must be at the core of interventions.

Fortunately, in recent years, research and practices have advanced. For example, we can adequately identify flood prone and vulnerable regions, putting us in a position to better plan ahead. Moreover, we have improved our capabilities to predict the occurrence and the impacts of a flood in near-real time, paving the way for the deployment of early warning systems. Additionally, we are in a better position to assess the performance of nature- and engineering-based interventions. While eliminating the risk of flooding completely is not possible, such interventions may complement a comprehensive strategy that aims in enhancing the resilience of communities to such risk.